Slots Autoplay Off by Default Cuts Loss-Chasing Top-Ups 19%
A two-year study of 41,600 U.S. online slot accounts found that turning autoplay off by default cut the share of sessions followed by an immediate deposit within 30 minutes — what the researchers call a "loss-chasing top-up" — by 19.2%, from 8.9% to 7.2%. The effect held across bet sizes and game types, and it showed up without any change to how much players could wager, how long they could play, or what they were allowed to deposit.
The study, run by a research group contracted by three licensed operators in Pennsylvania, Michigan, and New Jersey, is one of the first large-scale natural experiments on slot machine interface defaults in a regulated U.S. market. It lands as state regulators weigh whether default settings — not just warnings and limits — belong in responsible gambling rules, and as operators quietly test whether friction costs them revenue.
What the study actually measured
The setup was straightforward. In early 2023, the three operators agreed to flip the autoplay default on their slot lobbies from "on" to "off" for a randomly assigned half of new accounts. Existing accounts were assigned by the last digit of their account number. Nobody was told. Players in the control group saw the same spin button with autoplay available in a menu; players in the treatment group saw a single-spin button and had to navigate two screens to enable autoplay, with a prompt asking them to set a session loss limit first.
The researchers tracked four outcomes over 24 months: session length, average bet size, deposit frequency, and the specific behavior they labeled a loss-chasing top-up. That last one was defined narrowly. A top-up counted only if a player deposited within 30 minutes of a session ending, and only if that session ended with a net loss of at least 20% of the player's starting balance. Deposits after winning sessions, or after losses smaller than that threshold, were excluded. So were deposits made more than 30 minutes later, which the authors treated as ordinary reload behavior rather than an in-session reaction.
That narrow definition matters. A 19.2% relative reduction in a tightly defined behavior is not the same as a 19.2% drop in total deposits, and the study is careful not to claim that. Total deposits per player fell 6.1% in the treatment group, and net gaming revenue per player fell 4.4%. The gap between those numbers and the 19.2% figure is the story: autoplay defaults changed the timing and shape of deposits more than the total amount players were willing to spend.
The 30-minute window
The choice of a 30-minute window is the study's most contestable design decision, and the authors know it. In an appendix, they rerun the analysis at 10, 15, 60, and 120 minutes. The effect shrinks as the window widens — 22.4% at 10 minutes, 20.1% at 15, 14.8% at 60, 11.3% at 120. That gradient is consistent with the interpretation that autoplay specifically disrupts the fast loop between a losing session and a reload, rather than suppressing deposits generally. If the effect were just "players deposit less often," it wouldn't decay that way.
It's also consistent with a simpler explanation: players in the treatment group took longer to start their next session, so fewer of them happened to deposit inside any fixed window. The authors address this by controlling for time-to-next-session and find the effect persists, though attenuated to 15.6%. That's still a meaningful number, but it's the one I'd want to see replicated before treating 19.2% as the headline.
Why autoplay is the lever regulators keep circling
Autoplay has been the most-regulated feature in slot design for a decade, and the U.S. has been the outlier. The UK Gambling Commission banned autoplay on slots entirely in 2021, along with losses disguised as wins and several other features. Sweden's regulator capped autoplay at 25 spins. Australia banned it on pokies years earlier. In the U.S., autoplay remains legal in every state that offers online slots, and most operators ship it enabled by default.
The industry's argument for autoplay is mechanical: it reduces repetitive strain, it lets players set a spin count and walk away, and it's popular. That last point is real. In the study's control group, 38% of sessions used autoplay at least once, and among high-volume players — those betting more than $5 per spin — the figure was 61%. Autoplay is not a fringe feature.
But the study's most interesting finding isn't about use. It's about default. Among treatment-group players, autoplay usage fell to 11% of sessions. The feature was still available; it just wasn't pre-selected. That 27-point drop in usage produced a 19.2% drop in loss-chasing top-ups. The mechanism isn't that autoplay is dangerous per se — it's that a default is a decision, and most players don't revisit decisions.
This is the same logic behind default-enrollment retirement plans and organ donation opt-in rates. It's also why the finding is politically awkward. If defaults move behavior this much, then the industry's position that autoplay is a neutral convenience feature is hard to square with the data. The feature is neutral; the default is not.
The "losses disguised as wins" problem
There's a related design issue the study didn't test but that regulators often bundle with autoplay: losses disguised as wins, or LDWs. These are spins that return less than the amount wagered but trigger celebratory animations, sound, and sometimes a positive number on screen. On a $1 spin, a $0.40 return can look and sound like a win.
LDWs are legal in the U.S. and common. The UK banned them on slots in the same 2021 package as autoplay. The argument for banning them is that they distort a player's perception of whether they're winning, and autoplay compounds the distortion because the player isn't watching each spin. A player running 100 autoplay spins at $1 can lose $60 and see 30 "win" animations. The study's authors note that autoplay and LDWs likely interact, but they couldn't isolate the effect because all three operators used LDWs in both arms of the trial.
That's a real limitation. If the treatment effect is partly driven by players paying more attention to individual spins when they're clicking manually, then the 19.2% number is capturing something broader than the default itself. A follow-up trial with LDWs disabled in one arm would answer that, and none of the three operators has agreed to run one.
What operators say, and what the numbers say back
I asked all three participating operators for comment. Two declined. The third, a mid-sized operator licensed in Michigan and Pennsylvania that asked not to be named, gave a response that's worth quoting at length because it captures the industry's position precisely:
"Autoplay is a player-preference feature, and we support players choosing how they play. We participated in this research to understand the impact of interface changes, and we're reviewing the findings. We have no plans to change our default."
That's a defensible business position. A 4.4% drop in net gaming revenue per player is not nothing — for a mid-sized operator doing $200 million in annual slot revenue, it's roughly $8.8 million. But the arithmetic gets more complicated when you look at where the revenue comes from.
The study breaks revenue into deciles by player spend. The top 10% of players by spend accounted for 61% of net gaming revenue in the control group. In the treatment group, that share fell to 56%, and the top decile's loss-chasing top-ups fell 24.1% — higher than the overall effect. The bottom 50% of players by spend showed almost no change in top-up behavior, though their autoplay usage dropped just as much.
In other words, the default change hit the heaviest spenders hardest. That's the group operators are most reluctant to disturb, and it's also the group that responsible gambling researchers most want to reach. The study's lead author, who asked not to be named because the research is under peer review, put it bluntly: "The players whose behavior changed most are the players the feature was most likely to harm."
There's a counterargument, and it's not frivolous. If heavy spenders migrate to operators that keep autoplay on by default, the treatment operators lose revenue without reducing harm. Gambling is a partly substitutable market, and online slots are close to perfectly substitutable across licensed operators in the same state. A single-operator default change might just move the behavior.
The study can't rule this out. It tracked the treatment group's play within the participating operators only, and it didn't have access to data on whether those players opened accounts elsewhere. The authors flag this as the single biggest threat to the finding's policy relevance.
The multi-operator problem
This is the crux of the regulatory question. A default change is only as strong as its coverage. If Pennsylvania's Gaming Control Board mandated autoplay-off defaults for all licensed operators, players couldn't easily route around it — they'd have to leave the regulated market, which is a much higher-friction move than switching from one app to another.
The Pennsylvania board declined to comment on whether it's considering such a rule, but two people familiar with the board's internal discussions said staff have reviewed the study and that autoplay defaults are "on the list" of items under consideration for the next responsible gambling rulemaking cycle. The board's current slot rules date to 2021 and don't address interface defaults at all.
Michigan's regulator has been more explicit. In a September 2024 advisory bulletin, the Michigan Gaming Control Board noted that it was "monitoring research on default settings" and encouraged operators to "review whether pre-selected features align with the state's responsible gambling objectives." That's not a mandate, but it's a signal.
New Jersey, the largest of the three markets in the study, has said nothing publicly. Its Division of Gaming Enforcement has historically preferred to let operators self-regulate on interface issues, and it has been slow to adopt design-based rules even where other states have moved.
The compliance cost nobody's talking about
If autoplay-off defaults became mandatory, the operational cost would be modest. Operators would change a configuration flag and, in some cases, update a menu flow. The study's authors estimate the engineering cost at "under 40 hours per operator" based on the three participating platforms. That's not a barrier.
The real cost is competitive. An operator that turns autoplay off by default while its competitors keep it on is at a disadvantage in exactly the segment that drives revenue. That's a collective action problem, and it's why voluntary adoption has been slow. The study's authors note that two of the three participating operators reverted to autoplay-on defaults within six months of the trial ending. The third kept the treatment default in place but only for new accounts.
This is the pattern in iGaming responsible gambling generally: measures that work are often measures that individual operators can't afford to adopt alone. Deposit limits, session reminders, and now interface defaults all share this structure. The policy solution is a level playing field, which means regulation, which means the industry's trade groups will fight it.
The American Gaming Association's position is that responsible gambling tools should be "available, not mandatory," and that default settings are a matter of "player choice and operator discretion." That framing has held for years. Whether it holds against a 19.2% effect size in a peer-reviewed study is an open question — and it's the question that will decide whether autoplay defaults change in the U.S. anytime soon.
What to watch
Three things would move this from a single study to a regulatory consensus. First, replication in a market where autoplay is already off by default — the UK is the obvious candidate, though its data is harder to access. Second, a trial that isolates the LDW interaction, which would tell regulators whether they're looking at one design problem or two. Third, and most important, evidence on substitution: if players who lose autoplay defaults simply move to another operator, the case for a single-operator mandate collapses and the case for a statewide one gets much stronger.
None of those exist yet. In the meantime, the study's most durable contribution may be simpler than its headline number. It shows that a feature can be harmless in isolation and harmful as a default — and that the difference between those two things is a checkbox most players never touch.